Hormone Replacement Therapy and Lens Opacities: Title and subTitle BreakThe Salisbury Eye Evaluation Project FREE

CATARACT IS the leading cause of visual impairment in older adults in the world.¹ Improvements in living conditions, public health, and medicine have resulted in a dramatic increase in the number of people living past the age of 65 years, so cataract will continue to be a public health issue in the future. By 2020, an estimated 52 million Americans will be older than 65 years.² In the United States at present, more than 1 million cataract operations are performed annually at a cost of $3.4 billion, consuming 12% of the Medicare budget, and a steady increase is projected.³ Despite effective surgical treatment for the disease, cataracts place a large burden on the older population and the health care system. The identification of a factor to slow the progression of lens opacification would translate into a significant reduction in cataract surgeries and greatly improve the quality of life for older adults.

After the age of menopause, it has been consistently noted that women begin to have higher rates of cataract than men.^{4 - 6} Two large cross-sectional studies, the Beaver Dam Eye Study and the Blue Mountains Eye Study, have found protective associations between current hormone use and lens opacity. However, the Beaver Dam Eye Study⁷ found the association to be with nuclear opacity, whereas the Blue Mountains Eye Study⁸ found an association with cortical opacity and none with nuclear opacity. In both populations, women were included who had not reached menopause and whose endogenous estrogen exposure was not yet diminished. Recently, the Beaver Dam Eye Study⁹ published a report showing no association between use of hormone replacement therapy (HRT) and incident opacities for 5 years. However, it is unclear whether the study looked at current HRT use or duration of HRT use, and this study may not have had the power to detect an association.

Thus, some evidence exists that HRT may be associated with lens opacity, but there is no consistency about which type of opacity, nor has the role of endogenous and exogenous estrogen exposure been clarified. Data from a population-based study from Salisbury, Md, were used to determine whether use of HRT, after controlling for endogenous estrogen exposure, is associated with lower rates of lens opacities.

The Salisbury Eye Evaluation project identified a population-based random sample of elderly people aged 65 through 84 years living in Salisbury, Md, in 1993. Details about sample recruitment and methods are described elsewhere.¹⁰ A total of 65% completed the home questionnaire and the medical examination, resulting in 2520 participants. For this study, we excluded men (n = 1062), those with previous bilateral cataract surgery (n= 157), and those without photographs (n = 62), for a total of 1239 women with at least 1 phakic eye for whom photographs were graded.

Interviewer-administered questionnaires were given to participants at home. Information was obtained on use of HRT and birth control pills and reproductive and medical histories. Questions concerning HRT requested information about the type, dose, duration, and recency of use. Reproductive history questions requested information on age at menarche, age at menopause, pregnancy history, use of birth control pills, and hysterectomy. Medical history questions inquired about diabetes, hypertension, and previous eye surgery. Diabetes was evaluated on the basis of self-report or glycosylated hemoglobin level of at least 7.0%.¹¹ Hypertension was defined as systolic blood pressure of at least 160 mm Hg, diastolic blood pressure of at least 90 mm Hg, or self-report. Information was also collected on potential risk factors for cataract, including UV-B exposure, alcohol consumption, smoking, education, age, body mass index(BMI; calculated as weight in kilograms divided by the square of height in meters), and corticosteroid use.

Height and weight were measured by trained observers, as described by Caulfield et al.¹² Visual function was evaluated during a 4-hour eye examination performed at a Salisbury Eye Evaluation clinic as described by West et al.¹⁰ Pupils were dilated, and 2 nuclear photographs were obtained of each eye using a Topcon SL5D photographic slitlamp (Topcon Corporation, Tokyo, Japan) with the slit beam set at a height of 9 mm and a width of 0.1 mm and angled at 40°. Cortical photographs were obtained using a retroillumination camera (Neitz Instrument Company, Tokyo), focused just posterior to the papillary margin. A distance-recording device was set to 0 at this point. The focus was then placed at the posterior subcapsular (PSC) region. After the PSC photographs were obtained, the distance traveled to the back of the lens by the photographer was recorded. All photographs were processed by Wilmer Eye Institute Photography Service, Baltimore, Md, using standard processing techniques. Photographs of each eye were graded twice without knowledge of the status of the fellow eye. Photographs were graded for type and severity of opacity using the Wilmer grading scheme. Nuclear opacity was defined as a grade of 2 or higher on a 4-point continuous scale; cortical opacity, as a grade of 3 or higher on a 16-point continuous scale; and PSC, as present or absent. The eye with the maximum opacity grade was used in the analysis. If the 2 graders disagreed by more than 0.5 unit for nuclear or 1/16 unit for cortical opacity, an adjudicated grade was used. If an individual had a previous unilateral cataract surgery, data were used from the other eye. If the individual had previous bilateral cataract surgery, she was excluded from the analysis.

Individuals with and without cataract types were compared for different distributions of factors of interest. Present and past users of HRT and those who never used HRT were also compared for factors of interest. Since age is a major risk factor for cataract, we performed age-adjusted analyses for each factor to examine crude relationships with the different types of cataract. We then used multiple logistic regression to determine the associations between various estrogen factors and cataract type while simultaneously adjusting for other factors. The potential confounders included in these models were age (continuous), race, history of hypertension, history of diabetes, alcohol consumption (ever), smoking (current, past, or never), age at menopause, age at menarche, pregnancy (ever), number of births, corticosteroid use (ever), duration of corticosteroid use, hysterectomy (ever), BMI, education (≥12 years or <12 years), birth control use (ever), and duration of birth control use. These variables were chosen because they had different distributions within the exposure categories or had been shown in previous literature to be associated with cataract. In addition, cumulative UV-B exposure was included in models for cortical opacity, since it had previously been shown to be a risk factor. Stepwise selection procedures were then used to help choose the most parsimonious models. We evaluated biologically plausible interactions by stratification and by the addition of interaction terms in the model. All analyses were performed using a commercially available statistical package(SAS; SAS Institute Inc, Cary, NC).

Of the 1239 postmenopausal women aged 65 through 84 years in the final analyses, 398 (32%) had ever used HRT, including 134 (11%) who were using HRT at the time of the study.

As shown in Table 1, women who used HRT were more likely to be young, better educated, white, and lean; to have used medications such as birth control pills and corticosteroids; to have used alcohol or smoked in the past; and to have had a hysterectomy. They were less likely to have had conditions such as diabetes and hypertension. We compared women who were excluded (n = 219) with those who were included. Excluded women were more likely to be older and white and to use corticosteroids.

The age-adjusted associations between exogenous estrogen intake factors and opacity types are shown in Table 2. Current HRT use was associated with a lower prevalence of nuclear opacity(odds ratio [OR], 0.6; 95% confidence interval [CI], 0.4-1.0). In addition, HRT intake within the past 3 years was associated with decreased rates of nuclear opacity (OR, 0.6; 95% CI, 0.4-0.9). Duration of HRT use was not associated with nuclear opacity.

Association of current HRT use with cortical opacity was not statistically significant. However, the OR was less than 1 (OR, 0.6; 95% CI, 0.3-1.1). Duration of HRT use of more than 3 years was associated with a lower rate of cortical opacity (OR, 0.5; 95% CI, 0.3-0.8).

Past HRT use was associated with lower rates of PSC opacity (OR, 0.4; 95% CI, 0.2-1.0). No statistically significant association of use of birth control pills and duration of their use with any opacity type was found.

The age-adjusted analyses of endogenous estrogen factors and opacity are presented in Table 3. Duration of exposure to endogenous estrogen (age at menarche until age at menopause or hysterectomy) was not associated with any opacity type (data not shown). Women 16 years or older at first menarche had a higher prevalence of cortical opacity (OR, 1.7; 95% CI, 1.0-2.7). Age at menopause, hysterectomy, and number of births were not associated with any type of opacity in any consistent pattern, although later age at menopause showed a trend in the protective direction for nuclear opacity.

The comparison groups for each of the opacity types were allowed to contain the other opacity types. We also performed analyses using a comparison group excluding any other opacity, and the results were unchanged. In addition, we included an analysis for any opacity plus previous bilateral cataract surgery, and the results were unchanged (Table 2 and Table 3). The rates of current HRT use were lower in those who had previous bilateral cataract surgery (6%) compared with those who had any cataract (8%) or those who had no cataract (15%).

The results of the multivariate analyses were very similar to those of the age-adjusted analyses, as shown in Table 4. The associations of nuclear opacity with current HRT use(OR, 0.5; 95% CI, 0.3-0.7) and HRT use within the past 3 years (OR, 0.4; 95% CI, 0.3-0.7) were still present. For nuclear opacity, duration of use did not seem to matter, as both short- and long-term use were associated in the protective direction.

For cortical opacity, those who had taken HRT for a duration of more than 3 years had a lower prevalence of cortical opacity (OR, 0.6; 95% CI, 0.3-1.0). Current and recent use of HRT were not associated with cortical opacity. Age at menarche after multivariate adjustment was no longer associated with cortical opacity. Finally, we found an association of PSC opacity with current (OR, 0.3; 95% CI, 0.1-0.9) and past (OR, 0.2; 95% CI, 0.1-0.7) use of HRT.

In the multivariate analysis (Table 5) between current HRT use and nuclear opacity, markers of other estrogen exposure were also associated with lower rates of nuclear opacity. Women who had used birth control pills had a lower prevalence of opacity after adjusting for other factors (OR, 0.7; 95% CI, 0.4-1.0). Also, because estrogen may be stored in fat, the association of lower prevalence of nuclear opacity with a BMI of 25 to 30 (OR, 0.5; 95% CI, 0.4-0.7) or of greater than or equal to 30 (OR, 0.4; 95% CI, 0.3-0.6) compared with a BMI of less than 25 was of interest. This association was found previously by Caulfield et al.¹² In addition, an increasing number of births was associated with a lower prevalence of nuclear opacity (test of trend, P = .003). However, this association differed by age cohort. A quantitative interaction was found, with an increasing number of births being protective for those in the younger birth cohort (born 1918-1928) (OR, 0.9; test of trend, P = .05), whereas it was not associated with those in the older cohort (born 1908-1918) (OR, 1.0; test of trend, P = .61) (data not shown).

There was a statistically significant interaction between past use of HRT and pregnancy status (P = .03). Women who had never been pregnant had a lower prevalence of nuclear opacity with past (OR, 0.2; 95% CI, 0.1-0.7) and current (OR, 0.1; 95% CI, 0.0-0.6) HRT use. However, for women who had ever been pregnant, only current use was protective (OR, 0.5; 95% CI, 0.3-0.8). These data suggest that pregnancy itself, being a time of increased estrogen exposure, may confer a protective status, whereas past HRT use supplied protection for women who had never been pregnant. No statistically significant interactions were found with HRT use and BMI, use of birth control pills, hysterectomy, or smoking.

This study provides evidence that current and recent HRT use has a protective association against nuclear opacity, and that any HRT use has a protective association against PSC opacity, independent of endogenous estrogen exposure. This study also found markers of estrogen exposure, such as use of birth control pills and an increasing number of births, to be associated with a lower prevalence of nuclear opacity.

Our finding of an association of current HRT use with lower prevalence of nuclear opacity agrees with the results of the Beaver Dam Eye Study. That study did not find a statistically significant association between current HRT use and PSC opacity, although the OR was in the protective direction among older women. Their study also found lower rates of more severe nuclear opacity in the younger group of women who had used birth control pills.⁷ Our results do not support the findings from the Blue Mountains Eye Study, which reported lower rates of cortical opacity with current HRT use, higher rates of PSC opacity with current HRT use, and no association with nuclear opacity.⁸ The reasons for these differences are not obvious. The system for assessment of lens opacity in the Blue Mountains Study was similar to that in the Beaver Dam Study, whose results were similar to ours. Thus, differences in lens opacity grading systems are unlikely to be responsible.

An interaction between number of births and age was found with nuclear opacity. In the younger cohort, the more children a woman had, the lower the prevalence of nuclear opacity that was found after adjusting for other factors, including education and race. There was no protection with an increasing number of births in the older cohort. It is possible that aging effects overwhelm the protective effect of an increasing number of births.

Evidence of biological plausibility in the protection of estrogen against lens opacification can be found in the laboratory. A study in rats has shown that in a model of transforming growth factor β–induced cataractogenesis that is similar in morphologic and molecular biological features to human cataractogenesis, estrogen protected against the development of lens opacification.¹³ A similar study by Bigsby et al¹⁴ showed a protective effect of estrogen in a rat model of methylnitrosourea-induced cataract. Opacities developed after 8 months in only 12% of ovariectomized rats that received estradiol, whereas 74% of the control rats showed evidence of opacities.¹⁴ Furthermore, it has been demonstrated that estrogen receptor messenger RNA is present in the lens.^{15 - 16} In addition, a study¹⁷ investigated the effects of HRT on lens transmittance measured by means of fluorophotometry. The authors found less opacification among the postmenopausal women who had used HRT for at least 4 years compared with postmenopausal women who had not used HRT or men of the same age.¹⁷ Finally, a study has recently found that women receiving long-term tamoxifen citrate therapy, an antiestrogen, have an increased risk for cataract.¹⁸ Because users of HRT differ in many ways from nonusers, the associations might be due to other factors related to opacity development and HRT use. However, data were collected on many of the known potential cataract risk factors, and even after adjustment in the multivariate analyses, the associations remained relatively unchanged from the age-adjusted associations. In addition, the HRT users and nonusers had very similar numbers of overall comorbidities, eg, arthritis, hip fractures, heart problems, asthma, and cancer. Such data indicate that the lower prevalence of opacities in HRT users is unlikely to be related to a higher overall health status. Because of the multiple comparisons made in this study, some of the associations might be due to chance alone. However, the consistency of the negative associations across multiple HRT variables and the consistency of our results with those of other work support our findings as real.

Subjects with previous bilateral cataract surgery were excluded (n = 157) because we could not be certain of the time of surgery in relation to HRT exposure, and because we could not be sure as to the subtype and severity of opacity. These exclusions could have resulted in bias, although rerunning the analyses including these bilateral cataract surgical cases with those having any cataract yielded essentially the same results. Also, opacity information could not be obtained from some study participants because of absence of photographic data (n = 62). However, analyses of the characteristics of all the excluded participants (Table 1) indicated that they probably had higher rates of cataract because of older age and more corticosteroid use, and that they were less likely to have ever used HRT (28% vs 32%). Thus, including this group would likely serve to strengthen the protective associations of HRT use.

Our data suggest that the current and recent use of HRT may protect against nuclear and PSC opacification, the 2 most visually disabling cataract types. Prospective studies and clinical trials are needed to provide definitive answers, but in the context of previous studies and laboratory data, the results suggest that a potentially modifiable factor in cataractogenesis may be a woman's exposure to postmenopausal estrogen.

Accepted for publication February 23, 2001.

This work is supported by grant AG16294 from the National Institute on Aging, Bethesda, Md; a Research to Prevent Blindness Senior Scientific Investigator award, New York, NY (Dr West); and a grant K24 EY00395 from the National Eye Institute, Bethesda (Dr Schein).

Corresponding author and reprints: Sheila K. West, PhD, Wilmer Eye Institute, Room 129, Johns Hopkins Hospital, 600 N Wolfe St, Baltimore, MD 21210.